Grouting quality detection of prestressed tendon ducts using array ultrasonic tomography

Abstract

The grouting quality inspection of prestressed tendon ducts is crucial for detecting internal defects in prestressed structures. However, accurately assessing the grouting compactness within tendon ducts remains a significant challenge. To explore the feasibility and effectiveness of using array ultrasound for evaluating the grouting quality of prestressed ducts, this study proposes a method that leverages an improved deep learning model for identifying ducts and a technique based on ultrasound C-scan images for grouting quality assessment. Specifically, (1) by utilizing the proposed improved target detection model YOLOV7-SE, automatic recognition and localization of ducts in ultrasound images are achieved, which aids in selecting the appropriate depth range for C-scan image; (2) based on laboratory test results, criteria and procedures for assessing grouting quality using array ultrasound C-scan images are established. Ultrasound testing on simulated prestressed box beam top and web plates shows that the improved model accurately locates and marks the positions of tendon ducts, providing a reliable basis for setting the depth of ultrasound C-scan images. The reflective characteristics of the ultrasound C-scan images allow for a qualitative assessment of the grouting quality within the inspected area. The feasibility of the proposed method has been effectively validated through laboratory tests and field inspections.